Fullscreen HTML5Med Res (??MB)HTML4

Brilliant to the bone Techno Talk From bone-conduction headphones to exploring RealSense’s depth Max the Magnificent cameras and Eyesynth’s futuristic NIIRA headset, this month we journey into sound, vision and accessibility technology with life-changing potential. I mentioned this last month but, as a reminder, earlier this year the folks at H2O Audio (h2oaudio.com) gifted me a pair of their TR2 multisport, waterproof bone conduction beauties, which even work underwater if I want to listen to music while swimming. I must admit that before actually taking these for a spin, I had wondered about the quality of sound that can be achieved using bone conduction technology. I can honestly say that they exceed my expectations. To my untrained ear, the sound quality is as good as that offered by my regular headphones. Of course, they don’t block out external sound, but that’s an advantage in many situations. For example, when my wife (Gigi the Gorgeous) and I recently visited her mother, who lives in Nashville, Tennessee, I spent a delightful couple of hours trailing them as they meandered their way around a series of stores. I say “delightful” because I was happily enjoying a science fiction audiobook while still being able to detect and respond to the occasional dulcet tones directed at me by my spouse. They’re also great if you’re on a bicycle, scooter or similar, or even just walking around the streets, as you can still hear vehicles approaching, sirens and so on. Wearing them allows you to listen to music, audiobooks etc without ruining your situational awareness. Similarly, bone conduction technology offers an ideal solution for audio aids intended for the visually impaired, who rely on being able to hear what’s going on around them. We will return to this in a moment, but first… RealSense’s D415 A few days ago, I had a very interesting chat with the folks from RealSense (realsenseai.com). This used to be a division at Intel, but they recently spun out to begin operating as an independent company. Their primary focus (no pun intended) is computer vision technologies, particularly depth-sensing systems, that enable machines to perceive and understand their environment. One example of this is their D415 Depth Camera. There’s much more to this than meets the eye (OK, pun intended this time; sorry). Let’s start with the two RGB+ cameras mounted on the left and right sides (I’ll explain my use of the ‘+’ qualifier in a moment). These are used to provide stereoscopic depth perception in much the same way as human binocular vision. Each camera captures the same scene from slightly different angles. By identifying matching features in both images, the system measures their horizontal displacement, called disparity. Nearby objects show a significant disparity, while distant ones show little. Using triangulation, the system converts disparity into depth, producing a 3D depth map where each pixel encodes distance. This enables applications like object detection, robot navigation, augmented reality and gesture tracking. This is the clever bit. If we were to use regular RGB cameras, their performance would be degraded by poor lighting and low-texture surfaces, such as a flat white wall. The CMOS sensors used in RGB cameras are sensitive to a broad spectrum, including infrared (IR). If left unchecked, this IR can ‘bleed’ into the RGB channels, resulting in strange, washed-out colours. That’s why traditional RGB sensors include an IR-cut filter to remove the IR component from the image. The RGB+ cameras in the D415 don’t have this filter, meaning they also pick up infrared light, hence the ‘+’. Now observe the big ‘thing’ to the left of centre (I hope I’m not being too technical). This is an IR projector that casts a random pattern of thousands of dots, allowing the system to generate a 3D depth map even in a dark room and/or when viewing a flat, unicolored surface, such as a painted wall. The other ‘thing’ to the right of centre is a regular RGB camera. This captures a traditional RGB image of the scene. This image is 100% aligned with the 3D depth map, frame by frame and pixel by pixel. The D415 makes this data available for use by other people’s applications and systems (robotics, industrial automation, 3D scanning and modelling, medical imaging, motion tracking etc). Speaking of which… Eyesynth’s NIIRA One application that really (I’m going to resist saying “caught my eye”) attracted my attention is the NIIRA audiovisual perception headset from Eyesynth (eyesynth.com). If you look closely at this device, you’ll see it The D415 Depth Camera (Source: RealSense). 42 Practical Electronics | November | 2025 Techno Talk Max the Magnificent The NIIRA audiovisual perception device (Source: Eyesynth). has a RealSense D415 depth camera seamlessly moulded into the frame above the lenses. This headset connects to a small AI computer, roughly the size of a pack of playing cards. The RealSense D415 delivers the raw perception layer, and the Eyesynth NIIRA system adds the intelligence, interpreting the 3D scene, performing object detection, recognition, and contextual analysis. Any relevant information is translated into soundscapes that are conveyed to the user via the NIIRA bone-conduction audio system. As we previously noted, people with visual impairments rely heavily on sound cues from their surroundings. The last thing they want is to block their ears with headphones or diminish their hearing with earbuds, hence the use of bone conduction technology. A few years ago, I was invited to give a talk on advanced technologies at Sheffield Hallam University. Knowing Graham’s interest in this sort of thing, I invited him to attend. A lot has changed since I was a student there. I graduated in 1980. Goodness gracious me. I just realised that was 45 years ago. I’m too young for all this excitement! The thing is that I got lost on the way to my designated lecture theatre. I found myself standing at an intersection of corridors whose layout would have made the designer of the Paris street system blush with pride. I was confused, which is my natural state, so at least I was playing to my strengths. Then I heard a “tap tap tap” sound, and Graham appeared from around a corner, striding confidently along. I’m embarrassed to say that it was Graham who led us to our destination. If I ever win the lottery, I think it’s safe to say that Graham will find a NIIRA audiovisual perception device lurking in his Christmas stocking. PE Me aged ~3 (left) and Graham aged ~6 (right). My cousin Graham My cousin Graham is three years older than I am. That means he knew much more than I did when we were younger. I know this to be true because he told me so. He also told me that spaghetti was made from peeled worms, which has left a bad taste in my mouth to this day. About 30 years ago, Graham started to lose his sight. He’s now totally blind, but that hasn’t slowed him down. He danced with gusto and abandon at my wedding; he hosts a local radio show; and he often attends concerts in the city centre, making his way there and back again by bus and on foot. Practical Electronics | November | 2025 43